Silica-Supported Tungsten Carbynes (≡SiO) - ACS Publications

Feb 2, 2015 - Wei Bai , Wei Wei , Herman H. Y. Sung , Ian D. Williams , Zhenyang Lin ... Albert Poater , Ali Hamieh , Santosh Kavitake , Edy Abou-Hama...
0 downloads 0 Views 1MB Size
Article pubs.acs.org/Organometallics

Silica-Supported Tungsten Carbynes (SiO)xW(CH)(Me)y (x = 1, y = 2; x = 2, y = 1): New Efficient Catalysts for Alkyne Cyclotrimerization Nassima Riache,† Alexandre Dery,† Emmanuel Callens,*,† Albert Poater,†,‡ Manoja Samantaray,† Raju Dey,† Jinhua Hong,§ Kun Li,§ Luigi Cavallo,† and Jean-Marie Basset*,† †

Physical Sciences and Engineering, Kaust Catalysis Center, King Abdullah University of Science and Technology (KAUST), Thuwal 23955-6900, Saudi Arabia ‡ Institut de Química Computacional i Catàlisi and Departament de Química, Universitat de Girona, Campus de Motilivi, E-17071 Girona, Spain § Advanced Nanofabrication, Imaging and Characterization Core Laboratory, King Abdullah University of Science and Technology, Thuwal 23955-6900, Saudi Arabia S Supporting Information *

ABSTRACT: The activity of silica-supported tungsten carbyne complexes (SiO)xW(CH)(Me)y (x = 1, y = 2; x = 2, y = 1) toward alkynes is reported. We found that they are efficient precatalysts for terminal alkyne cyclotrimerization with high TONs. We also demonstrate that this catalyst species is active for alkyne cyclotrimerization without the formation of significant alkyne metathesis products. Additional DFT calculations highlight the importance of the W coordination sphere in supporting this experimental behavior.



INTRODUCTION Cyclotrimerization [2 + 2 + 2] of alkynes catalyzed by transition metals is an elegant chemical transformation for the single-step construction of multiply substituted aromatic compounds. This reaction, discovered by Berthelot in 1866, refers to the addition of three acetylene molecules to form benzene derivatives at very high temperatures.1 Later on, Reppe found that the transition-metal complex Ni(CO)2(PPh3)2 could efficiently catalyze this reaction.2 Since then, numerous procedures employing late transition metals rather than earlier transition metals have been reported for this alkyne cyclotrimerization.3−5 In particular, the use of group VI precursors in their high oxidation state is scarce. In the late 1960s, Boelhouwer and co-workers6 reported the cyclotrimerization of alkyne catalyzed by silica-supported tungsten trioxide, which was originally the catalyst used by Bailey and Banks for the first report on alkyne metathesis.6,7 In a fixed bed reactor, 1-pentyne (WHSV, 1−4 h−1) is transformed into tripropylbenzene derivatives at high temperatures (>250 °C). However, no mechanism was proposed to elucidate how the d0 WVI species in its high oxidation state could be active for alkyne cyclotrimerization. Thus, the late development of surface organometallic chemistry, in which an organometallic precursor is grafted in a controlled manner at the surface of an oxide, proved to be an ideal tool to reach a better fundamental understanding of heterogeneous catalysis.8 For instance, the design and synthesis of numerous well-defined silica-supported © XXXX American Chemical Society

tungsten oxo alkyl precursor complexes gave more insights to improve the catalyst performance for olefin metathesis.9 Ensuing this strategy, we reported recently the characterization of supported Si−O−WMe5 (1), which upon thermal treatment at ca. 150 °C evolves to the unprecedented W methylidyne/methyl species 2 (Scheme 1).10 Scheme 1. Transformation of Si−O−WMe5 (1) into ( SiO)xW(CH)(Me)y (2: x = 1, y = 2; x = 2, y = 1)

Following the isolation of this thermally stable W species 2, we decided to examine its activity toward alkynes. In our recent work, we observed an equilibrium between 2 and W bismethylidene promoted by coordination of either PMe3 or an olefin to 2.11 Moreover, it was demonstrated earlier that the use of bis-alkoxide alkylidyne species instead of a bis-alkyl alkylidyne is required to achieve alkyne metathesis, highlighting the importance of the W coordination sphere.12 Herein we Received: June 28, 2014

A

DOI: 10.1021/om500684e Organometallics XXXX, XXX, XXX−XXX

Article

Organometallics

that an initial alkyne metathesis produced 1-hexyne, which after cyclotrimerization give the aromatic products 3 (Scheme 3ii). The catalytic run with a mixture of 1-hexyne/5-decyne gave the same products as with the 1-hexyne experiment: tri- and tetrasubstituted benzene derivatives 3 and 4. Nevertheless, the tetrasubstituted product represents more than 13% of the overall products in comparison to 1.5% when using only 1hexyne. This result strongly supports that the cross-alkyne cyclization between 1-hexyne and 5-decyne is taking place, leading to the two observed products 4. Since both precursors 1 and 2 produce similar results for the 1-hexyne cyclotrimerization, we chose to employ 1 to examine other terminal alkynes (Table 1).

report the remarkable activity of supported complexes 1 and 2 for the alkyne cyclotrimerization and provide the origin of such reactivity toward alkyne metathesis by DFT calculations.



RESULTS AND DISCUSSION In a sealed tube, a catalytic amount of 1 or 2 was allowed to react with 1-hexyne at 150 °C. Cyclotrimerization of 1-hexyne catalyzed by either 1 or 2 gave a mixture of the two aromatic regioisomers 1,3,5-tri-n-butylbenzene (3a) and 1,2,4-tri-nbutylbenzene (3b) in quantitative yield with a molar ratio of 29:71, respectively (Scheme 2). Scheme 2. 1-Hexyne Cyclotrimerization Catalyzed by 1 or 2

Table 1. Alkyne Cyclotrimerization using 1a

The structure of these regioisomers was further confirmed by NMR spectroscopy. In 1H NMR spectroscopy, the signal at 6.82 ppm corresponds to the three equivalent protons bonded to sp2 carbons of the aromatic rings of 3a, while the set of signals at 6.93−6.95 and 7.05 ppm accounts for the CH aromatic resonances of 3b. GC-MS and high-resolution mass spectroscopy further confirmed the molecular weight of these aromatic compounds. No polymeric products were observed at the end of the catalytic run, as determined by the absence of proton resonances between 5 and 6 ppm. Furthermore, aromatic products 4 with an electron ionization fragment of m/z 302 were also observed (molar fraction 99%) (Figure 1). These supported W surface species were found to be highly active precursor catalysts for the 1-octyne cyclotrimerization with an initial TOF of 1.03 s−1 (measured at 15 min). One advantage of such system in comparison to the reported WO3/SiO2 is the lower temperature of the reaction (150 °C vs >250 °C). In this supported catalytic system, we assume that these W species have a “single-site” character. Knowing the propensity of these W methylidyne/methyl species to form bimetallic species in solution,14 we also studied the imaging characterization of oxidized supported Si−O−WMe5 to determine the distribution of the W atom on the surface by scanning transmission electron microscopy (STEM) with spherical aberration correction. The microscope was operated at an accelerating voltage of 300 keV, offering a high enough spacial resolution for atomic high-angle annular dark-field imaging

Scheme 3. Proposed Mechanism for the Formation of (i) 5Decyne from 1-Hexyne and (ii) 1-Hexyne from 5-Decyne

B

DOI: 10.1021/om500684e Organometallics XXXX, XXX, XXX−XXX

Article

Organometallics

were observed, supporting the notion that immobilized W atoms are well dispersed after the grafting of initial tungsten hexamethyl upon silica.15 Overall, the above results show that single-site species 1 and 2 efficiently catalyze terminal alkyne cyclotrimerization rather than the classical alkyne metathesis.16 Moreover, the alkyne substrate, which could act as a four-electron donor (σ + π), does not prevent the cyclotrimerization reaction, as is known with tris(alkyne) tungsten 18-electron complexes.17 This raises the question of the origin of such reactivity. DFT calculations were performed to elucidate the matter. We chose propyne as a substrate to avoid the complication of the conformational flexibility of the saturated n-C6 tail of 1-octyne. The process starts with the W−carbyne bipodal species 2, followed by coordination of the first propyne to form the tungstenacyclobutadiene derivative 10. The well-known loss of a proton from this metallacycle to the methyl group bound to W, releasing a methane molecule,18 is not favored in our case, since it is predicted to have a barrier of 69.7 kcal/mol. This is in agreement with the fact that no polymeric product is observed, since deprotonation of the metallacyclobutadiene would lead to a polymerization process.19 Rather, the further insertion of an alkyne molecule to give the WVI−cyclohexatriene 11 is less demanding, with a ΔG⧧ value of 16.8 kcal/mol, and is strongly exergonic, with an energy gain of 41.1 kcal/mol. Next, the WVI species 11 is reduced by overcoming a barrier of 11.8 kcal/mol to produce the bis-methylcyclopentadienyl WIV species 12, with an energy gain of 19.7 kcal/mol. This metallocene derivative precedes the third insertion, which starts with the easy alkyne coordination to the W center of 13, by overcoming a low barrier of 7.7 kcal/mol. Intermediate 13 evolves into 14 through formation of a C−C bond between the coordinated alkyne and a substituted C atom of the five-carbon ring. Although this step involves a significant barrier of 31.4 kcal/ mol, the alternative methane formation due to the interaction of the methyl group bonded to W and the nearest H atom of the five-membered ring costs 60.7 kcal/mol. Intermediate 14 evolves first to the bicyclo intermediate 15 and then to intermediate 16, presenting a seven-carbon ring and overcoming the high barrier of 29.6 kcal/mol. However, the formation of 16 is thermodynamically highly favored, by 30.5 kcal/mol, and this high stability of 16 can be also related to the next barrier of 35.0 kcal/mol to give species 17, which presents the final six-membered carbon ring. This step is considered as the rate-determining step of the reaction pathway. Finally the assistance of a fourth alkyne insertion leads to intermediate 17, overcoming a barrier of just 13.0 kcal/mol and facilitating the next release of the 1,2,4-methylbenzene. For this final step to regenerate the propagative species 10 to close the catalytic cycle, a 2 + 2 cycloreversion to release the trisubstituted benzene product overcomes the last barrier of 17.7 kcal/mol. To switch to the alternative formation of a 1,3,5trisubstituted benzene derivative, from species 15 a barrier 3.2 kcal/mol higher leads to the isomer 16′ with a symmetric 1,3,5-methyl substitution. The next steps are identical with those in the formation of the unsymmetrical product. On the other hand, the opposite regiochemistry of addition of propyne to the W−carbyne species 2, leading to the metallacycle 10″, was examined in Scheme S1 (Supporting Information). Formation of 10″ costs only 2.1 kcal/mol more than formation of 10. However, the 1,2,4-trisubstituted benzene formation is favored energetically as well. The formation of the metallocene intermediate 12 is in competition with an alkyne

Figure 1. Cyclotrimerization of 1-octyne catalyzed by 1. Experimental conditions: 1 (25 mg, 0.75 μmol, W loading 0.55 wt %, 1 equiv), 1octyne (3.38 mmol, 4500 equiv), 150 °C, batch reactor.

(HAADF), which is convenient to discriminate heavy-element atoms such as W. Since our complex is highly air-sensitive, a pressure of dry oxygen (1 atm) was introduced in a batch reactor and maintained for 10 min prior to analysis. With a high ratio C/W of the resulting powder, we also did a plasma oxygen cleaning to prevent the formation of carbon during the analysis. Figure 2a shows the atomic resolution HAADF image of

Figure 2. (a) HAADF STEM image and (b) HAADF intensity profile of oxidized species 1.

discrete single W atoms on the surface of silica. The bright spots correspond to single W atoms, since HAADF is a Z contrast imaging method. Figure 2b represents the HAADF intensity profile along the line imposed on Figure 2a. The dispersed peaks indicate that the heavy W atoms with large atom number Z = 74 and demonstrate a full width at halfmaximum (fwhm) of 0.135 nm, close to the W atom radius (0.141 nm). Therefore, single W atoms of oxidized species 1 C

DOI: 10.1021/om500684e Organometallics XXXX, XXX, XXX−XXX

Article

Organometallics Scheme 4. Proposed Mechanism for the Formation of 1,2,4- and 1,3,5-Trisubstituted Benzene Derivativesa

a

Energies are given in kcal/mol; carbon atoms are labeled in green.

is generally required in the coordination sphere of the W metal to favor the alkyne metathesis.20 Finally the alkyne addition step 10 to 11 was compared with release of acetylene from 10, through a classic alkyne metathesis step. We found the latter to be disfavored by 9.7 kcal/mol, explaining why alkyne cyclotrimerization is favored over alkyne metathesis under our reaction conditions. As final remark, we note that in many cases we did not investigate the regiochemistry of alkyne reactivity or other types of skeletal rearrangements. However, the number of possibilities to explore is practically infinite, which makes it impossible to investigate all of them in detail. Nevertheless, we believe that the overall catalytic cycle shown in Scheme 4 captures well the complex transformations promoted by this supported W catalyst, offering a sound rationalization of the experimental results.

insertion (see Scheme S2 (Supporting Information)). Indeed, this third insertion has two steps, consisting of a classic alkyne metathesis step leading to intermediate 18, with an energy barrier of 28.7 kcal/mol, followed by a rearrangement of the carbon skeleton to intermediate 19, with an energy barrier of 17.3 kcal/mol and an overall energy gain of 17.4 kcal/mol. This step can be seen as a Cope-like reaction, with the CC double bond of the metallacyclobutane acting as the olefin. Interestingly, the final intermediate 19 has a structure consistent with a metallacyclopentene, with two W−C σ bonds (W−C3 and W−C7) and with the sequence of bond lengths C3−C2 = 1.44 Å, C2−C1 = 1.41 Å, and C1−C7 = 1.44 Å, which is a long−short−long rather than a short−long−short sequence, which would have been consistent with a η4coordinated diene geometry. A final fourth insertion step of a propyne unit leads to intermediate 20, from which the 1,2,4trisubstituted benzene product is eliminated via an energy barrier of 27.8 kcal/mol, closing the catalytic cycle through recovery of the tungstenacyclobutene intermediate 10. Interestingly, intermediate 20 presents two clear σ bonds, W−C2 = 2.22 Å and C5−C6 = 1.54 Å, and two C−C double bonds, C1−C7 = 1.38 Å and C3−C4 = 1.36 Å. Bearing in mind that the consecutive insertion of alkyne molecules in Schemes 4 and Schemes S1−S3 (Supporting Information) is consistent with the low steric hindrance of these systems, this alternative mechanism favoring the third alkyne insertion instead of the formation of the five-carbon ring in species 12 cannot be excluded. Indeed, species 12 could be a stalemate in such an alternative mechanism. This consecutive alkyne insertion explains why the use of bulky alkoxide groups



CONCLUSION We reported an efficient cyclotrimerization of terminal alkynes. In addition, DFT calculations show why this chemical transformation is favored instead of the typical alkyne metathesis. This study clearly highlights the importance of the coordination sphere of the W metallacyclobutadiene which is detrimental for alkyne metathesis.



EXPERIMENTAL SECTION

All experiments were conducted under an argon atmosphere using glovebox techniques. The syntheses and the treatments of the surface species were carried out using high-vacuum lines (